Thinking Machines Lab co-founder Lilian Weng is returning to OpenAI after announcing earlier this week she was leaving TML over workload-related health concerns
The move follows reported departures from the young lab, including other employees expected to join OpenAI and Andrew Tulloch’s move to Meta. It makes workload sustainability part of the competitive talent story, not just an internal staffing issue.
First-order effects
OpenAI regains Lilian Weng, while Thinking Machines Lab loses a co-founder shortly after she cited workload-related health concerns in announcing her exit.
The return adds to visible personnel churn around Thinking Machines Lab, putting immediate pressure on the startup’s remaining leadership and research organization.
Second-order effects
Thinking Machines Lab must compete harder to retain and recruit researchers when its former employer can reabsorb senior founders; rival labs can use comparable stability and workload claims in recruiting.
For OpenAI, repeated inflows from the startup can strengthen its research bench, while concentrating attention on whether high-intensity frontier-lab work is sustainable for key staff.
Third-order effects
If departures from newer frontier labs continue to flow back to established platforms, independent labs may find founder-led research organizations harder to sustain through talent alone, reinforcing the earlier reported OpenAI hiring pipeline.
The episode points to frontier AI becoming more institutionalized: organizational support, management capacity and durable research environments may increasingly matter alongside technical ambition and compensation.
The trend: Frontier-AI competition is increasingly being decided by the ability of labs to retain elite researchers under sustained organizational and workload pressure.
It is a hard and sad decision. I shared this message with folks at Thinky. Thank you all for the time together♥️ Just as the last sentence in my message: The future worth building is human. [image]
your blog taught a generation of us how to think about RL, agents, and safety long before these became mainstream topics. thank you for everything youve given this field, and for the honesty of this note. health first, always. AGI is a marathon and the field will be here whenever…
Lilian Weng, the Thinking Machines cofounder who announced her departure earlier this week, citing startup-related stress and illness, is rejoining OpenAI. …
AI's next casualty might be the health of the people building it. — Thinking Machines Lab cofounder Lilian Weng said she's stepping down from her role because of stress and illness. …
Lilian Weng, co-founder of Thinking Machines, shared a heartfelt message to the company about leaving to focus on her health given constant stress and the workload. — Today OpenAI announced that Lilian Weng was rejoining them from Thinking Machines to lead an OpenAI research te…
Well this was strangely misleading. Thinking Machines cofounder posted honestly about stepping down due to overly fast pace. I assumed this was to take a break. Turns out just jumped ship to go back to work at OpenAI - where the pace is also anything but slow. So strange.
Hearing that Weng has already started at openai in her departure note, Weng hinted at where she was headed: “a scoped role in a more predictable place, and not as a cofounder, will suit the next stage of my life better.”
Lilian Weng is returning to OpenAI to work on recursive self-improvement: using AI models to help build better AI models. Looks like OpenAI is really bullish for RSI.
“you're putting your health first” is brutal female coded language for “coward, you abandoned the project because couldn't handle the pressure and I am really f angry at you” [image]
This is the weirdest industry. So many AI cofounders are willing to jump ship, change teams, sell fast, etc ... Never been true in past startup up-cycles.
The Thinking Machine co-founder posted a long diatribe about how her health was struggling due to work, so she needed to step back. Turns out she was just leaving to go to OpenAI lol
You have to understand when you get lucky in life. Hard work is not enough for success alone. The model lab researchers from OpenAI got supremely lucky. They thought that would mean anything else they did would succeed as well. Hard lesson to learn..
this is so bizarre! either thinky's culture is so toxic one must lie about their health to leave or openai is a significantly less intense place than thinky, which seems like a negative signal from a frontier lab? am i missing something?